1. MED. MCROBL.-VOL. 16 (1983) 117-127 1983 The Pathological Society of Great Britain and reland GENTAMCN AND METHCLLN RESSTANT STAPHYLOCOCCUS A UREUS N DUBLN HOSPTALS: CLNCAL AND LABORATORY STUDES MARY T. CAFFERKEY, ROSEMARY HONE,* F. R. FALKNER, c. T. KEANE AND HARRET POMEROY? Department of Clinical Microbiology, Trinity College, St James s Hospital, James s Street, Dublin 8, *Department of Microbiology, The Mater Misericordiae Hospital, Dublin 9 and?department of Microbiology, Moyne nstitute, Trinity College, Dublin 2 SUMMARY. Strains of Staphylococcus aureus resistant to gentamicin and methicillin first appeared in Dublin hospitals in 1976, and rapidly became widely disseminated. The number of patients infected or colonised increased throughout the period of study, especially in 1979 and 198. Most isolates were from burns, surgical wounds and traumatic skin lesions. During the 12 months after first isolation of these multiply antibiotic resistant strains, colonisation or minor infection was the usual event. nvasive infection such as bacteraemia, deep wound sepsis and osteomyelitis was rarely seen. Subsequently, as the number of patients from whom these organisms were isolated increased, bacteraemia and other severe infection became more common. The predominant phage type of S. aureus changed with the progression of the outbreak. solates of different phage type were sometimes found in a single lesion, or in different sites in one patient. By the second half of 198, most isolates were untypable or typed only with an experimental phage. NTRODUCTON Gentamicin resistance in a clinical isolate of Staphylococcus aureus was first reported by Lacey and Mitchell (1969). Such resistance was very rare until 1975. However, as a result of the emergence of neomycin resistance in S. aureus in association with the topical use of neomycin, Lacey (1975) predicted that gentamicin resistance would appear in S. aureus. This prediction became a reality in 1975 and 1976 with the appearance of S. aureus strains resistant to gentamicin and other aminoglycosides in single unrelated cases and in outbreaks of hospital infection in many centres (Soussy et al., 1975; Speller et al., 1976; Wyatt et al., 1977; Greenhood et al., 1979). A more serious development was the appearance of strains of S. aureus resistant to gentamicin and several other antibiotics including methicillin (Soussy et al., 1976; Shanson et al., 1976; Crossley et al., 1979; Price et al., 198). Received 1st June 1982; revised version accepted 19 Oct. 1982. 17
118 MARY T. CAFFERKEY ET AL. n the year from April 1971, 5.4% of S. aureus strains isolated from eight Dublin hospitals served by this laboratory were resistant to methicillin (Hone and Keane, 1974). These isolates were also resistant to most of the commonly used antibacterial agents, but all were sensitive to chloramphenicol and gentamicin. Phage typing showed that, in spite of varied patterns, these strains belonged to phage group 111; most were sensitive to phages 77 and 84. A gentamicin and methicillin resistant strain of S. aureus (GMRSA) was first isolated from a clinical specimen taken in one of our hospitals in July 1976. This strain was lysed by phages of groups and 111. Subsequently the frequency of isolation of such strains increased and bacteraemias and other invasive infections with these organisms have been observed. Most of these later strains were sensitive to phages of group 111, especially 77 and 84. Strains of this phage type have been isolated consistently from these hospitals since then. We report here the distribution of GMRSA in eight hospitals, the types of specimens from which these organisms were isolated and the phage-typing patterns of GMRSA isolates. The period of study was 1 Jul. 1976 to 31 Dec. 198. MATERALS AND METHODS The Hospitals. Eight Dublin teaching hospitals participated in this study. These were: (A) The Federated Dublin Voluntary Hospitals (F.D.V.H.), a group of seven hospitals totalling 1231 beds including the National Burns and Plastic Surgery Units and served by a central microbiology laboratory, and (B) the Mater Misericordiae Hospital (M.M.H.), with 449 beds and including the National Cardiac Surgery Unit. n this study the hospitals are designated numbers one to eight. S. aureus strains. S. aureus strains were identified by the tube coagulase test with human plasma (standardised normal plasma; Dade Diagnostics nc., Miami, FA). The Oxford strain of S. aureus NCTC 6571 was the control organism in antibiotic susceptibility tests. Because of the large number of gentamicin and methicillin resistant strains of S. aureus isolated, only selected representative strains were stored for further study; these included all isolates from invasive infection, sites normally sterile and epidemiological specimens and isolates from a single specimen showing differences in cultural characteristics or antibiotic susceptibility. solates selected for plasmid studies were stored in 4% glycerol at - 2 C; other isolates were stored on nutrient agar slopes at room temperature. Antibiotic Susceptibility Tests. Antibiotic susceptibility tests were performed by the Stokes disk diffusion method (Stokes and Waterworth, 1972) on Diagnostic Sensitivity Test Agar (DST, Oxoid Ltd) with the following disks: penicillin G 2 units, tetracycline 1 pg, erythromycin 15 pg, trimethoprim 1.25 pg, sulphamethoxazole 1 pg, gentamicin 1 pg, amikacin 3 pg, neomycin 3 pg, kanamycin 3 pg, tobramycin 3 pg, chloramphenicol3 pg, clindamycin 2 pg, fusidic acid 1 pg, rifampicin 3 pg, and vancomycin 3 pg. Single disks and Mastrings (Mast Laboratories Ltd, Liverpool) were used. The plates were incubated overnight at 37 C. Methicillin resistance was tested at 3 C using a 1 pg methicillin disk (Annear, 1968). Minimum inhibitory concentrations (MCs) of gentamicin, neomycin, kanamycin, tobramycin, amikacin and vancomycin were performed by the agar dilution method on DST agar. A multipoint inoculator was used to deliver 14-15 cfu of an overnight broth culture; this allowed 27 strains to be inoculated on a plate. Minimum bactericidal concentrations (MBCs) of gentamicin, kanamycin and amikacin, were determined by the tube dilution method. MCs for methicillin were determined by the tube dilution method incubated overnight at 3 C. Bacteriophage Typing. This was performed on all stored isolates. The method used was that of Blair and Williams (1961) using the routine set of phages together with the experimental phages 88A, 89 and 9. These, and their propagating strains were kindly supplied by the Director of the Staphylococcal Reference Laboratory, Colindale, London.
RESSTANT S. AUREUS N DUBLN 119 RESULTS solation of GMRSA The frequency of isolation of GMRSA from clinical specimens, or the proportion of all S. aureus isolates that were gentamicin and methicillin resistant was not known for all eight hospitals because records were kept of selected S. aureus isolates only. Every patient from whom GMRSA was isolated, and the affected sites, was recorded. However, for gentamicin-sensitive S. aureus, records were kept only of patients in whom the organism had caused significant infection or who were bacteraemic, and patients from whom S. aureus was isolated from the nose before cardiac surgery. GMRSA strains were isolated from clinical specimens from 849 patients. Fig. 1 shows the numbers of patients from whom GMRSA were isolated in each 3-month period. GMRSA were isolated from some patients for several months and, from seven patients, were isolated from an infected site for more than 12 months. The total of 849 is certainly an underestimate of the number of patients infected or colonised, as the indications for taking specimens varied greatly within and between hospitals. Fig. 1 also shows episodes of bacteraemia and other severe infection for each 3-month FG. 1. Numbers of patients from whom GMRSA were isolated in each 3-month period from 1 Jul. 1976 31 Dec. 198. -=Total patients, ---=patients with bacteraemia or other severe infections. to
12 MARY T. CAFFERKEY ET AL. period throughout the study. The incidence of bacteraemia and other severe infection increased especially in 1979. Spread of GMRSA in the hospitals The numbers of patients in the different hospitals from whom GMRSA were isolated during each 3-month period are shown in fig. 2. The number of patients so- 4. 3. 2. m - 6 5 4 3 2 ro n lo. FG. 2. Numbers of patients in the different hospitals from whom GMRSA were isolated in each 3-month period from 1 Jul. 1976 to 31 Dec. 198.
RESSTANT S. AUREUS N DUBLN 121 infected or colonised varied greatly between the different hospitals, but, in all cases showed a progressive increase during the study. All isolates in 1976 were from in-patients in Hospital 7. This hospital includes the National Burns and Plastic Surgery Units. Throughout the study, 44 of the 849 patients from whom GMRSA were isolated were in Hospital 7. Strains of GMRSA were isolated from only three patients in Hospital 4 and from only eight patients in Hospital 5. Episodes of bacteraemia and other severe infections were uncommon in Hospital 7 when the total number of patients colonised or infected is taken into account. The greatest incidence of severe infections was in Hospital 3. Relationship of GMRSA to duration of hospital stay. This was not studied formally. t was, however, noted that in burns patients, swabs taken at admission were usually negative for GMRSA; swabs taken subsequently (48 or 72 h later) frequently yielded GMRSA and the likelihood that the lesions would be colonised or infected with these organisms increased with duration of hospital stay. Sites of GMRSA infections Sites from which GMRSA was isolated have been divided into two groups-superficial and deep. Specimens from nose, throat, perineum, axilla and hair, which were considered to be carrier sites, were excluded. Tables and 1 show the numbers of patients from whom GMRSA were isolated in specimens from individual sites. Patients from whom GMRSA were isolated from more than one similar site are included once only, even though GMRSA isolates from these sites sometimes showed minor differences in morphology, antibiotic susceptibility pattern or phage type. When the results were analysed by these methods, 849 patients were colonised or infected at 924 sites of which 68 sites (73.6%) were superficial. Wounds-surgical and non-surgical including pressure sores and ulcers-and burns, were the sites from which GMRSA were isolated most frequently. The urinary TABLE solation of GMRSA from superficial sites of 849 patients (1976-198) Number of GMRSA strains isolated* in Site ofinfection 1976 1977 1978 1979 198 Total Wound Burn Foreign body? Ulcer Pressure sores Sinus Tracheostomy Conjunctivitis Arterial fistula Total 7 28 72 118 174 399 13 25 28 36 6 162 2 3 1 18 33 1 3 11 2 35 4 8 15 27 1 3 5 9 1 1 1 6 9 2 3 5 1 1 2 57 114 187 32 68 * One isolate only was recorded per site per patient; patients were sometimes colonised at more than one site. t Site of drain, drain tip, fluid from drain, pin, rod, pacemaker wire.
122 MARY T. CAFFERKEY ET AL. TABLE 1 solation of GMRSA from deep sites in 849 patients (1976-198) Number of GMRSA strains isolated* in - - Site of infection 1976 1977 1978 1979 198 Totals Urinarv tract uriie Catheter tip A} 1 :} 6 i} 17 2:} 26 :} 58 Ti} 18 SDutum 1 7 17 2 29 74 Blood 1 4 11 15 31 Abscess 1 3 6 1 Osteom yelitis? 1 8 9 Pleural fluid 4 2 6 Total 3 13 39 65 118 238 *One isolate only was recorded per site per patient; patients were sometimes colonised at more than one site. t All were post-operative infections and five were also post-traumatic. tract was the site of infection in 12.7% of patients; the organisms were isolated from sputum or tracheal aspirate from 8.7%. Bacteraemia was documented in 3.5% of infected patients. Presence of other organisms Superfzciul sites. GMRSA isolated from burns and graft wounds were usually in mixed culture, most commonly with various gram-negative bacilli. nfected surgical wounds usually gave a pure growth of GMRSA, and the appearance of the gram-stained smears was compatible with the presence of staphylococci only. The majority ( > 9%) of specimens from the sites of foreign body implants, drains and sinuses yielded pure cultures as did specimens from conjunctivae, arterial fistulas and intravenous infusion sites. n specimens from pressure sores and ulcers, a mixed flora was usually demonstrated by both gram-stained smear and culture; however, GMRSA was the predominant organism. Deep sites. A pure growth of GMRSA was isolated from urinary catheter tips, intravenous catheter tips, abscesses, pleural fluid and specimens from osteomyelitis. Gram-stained smears of the specimens revealed pus cells and gram-positive cocci in clusters. GMRSA isolates from urine were obtained in pure culture in c. 9% of cases. Sputum was cultured by a dilution method; most specimens yielded > lo7 colonies of GMRSA/ml. n all cases, the gram-stained smear of the specimen showed profuse pus cells and intra- and extra-cellular gram-positive cocci in clusters. solation of GMRSA from out-patients GMRSA were frequently isolated from infected patients for many months after discharge from hospital. This was not studied formally and isolates were from specimens taken on attendance at the out-patient or casualty department. Specimens from decubitus ulcers, pressure sores, operative wounds, stitch sinuses or wounds
RESSTANT S. AUREUS N DUBLN 123 TABLE 11 Antibiotic sensitivity of 58 GMRSA isolates from patients with severe infections Antibiotic Percentage sensitivity An ti bio tic Percentage sensitivity Penicillin Methicillin Erythromycin Clindamycin Fusidic Acid Tetracycline Kanamycin Gentamicin 24 44 6 Tobramycin Neom ycin Amikacin Sulphonamides Trimethoprim Chloramphenicol Ri fampicin Vancomycin * 56 92 Variable Variable 92 1 1 overlying irradiated areas accounted for the great majority of isolates. The S. aureus carriage sites were not studied in these patients. One specimen received from a general practitioner yielded GMRSA. The patient was a 3-year-old child with burns who attended two children's hospital casualty departments for dressings. The burns were being treated with topical fusidic acid. Antimicrobial sensitivity tests Results of disk tests with 59 strains of GMRSA are shown in table 111. All strains were sensitive to rifampicin and vancomycin. Most were resistant to sulphonamides (Su) and trimethoprim (Tmp). However, Su and Tmp sensitivities were variable; of otherwise indistinguishable colonies picked from a seemingly pure strain, some were resistant while others were sensitive to these two antibiotics. When Su and Tmp susceptibility testing was performed on isolates from successive specimens from the same site, similar variation was seen. Fig. 3 shows the MCs of gentamicin, tobramycin, kanamycin, neomycin, 1 C 8.-.- c 1 // / / i L a" 2.5 1 2 4 8 16 32 64 128 286 S12 Concentration of antibiotic (m g/l) FG. 3. Percentage inhibition of 33 strains of S. aurew by increasing concentrations of antibiotic. = Amikacin, = neomycin, = gentamicin, = vancomycin, A = tobramycin, A = kanamycin.
124 MARY T. CAFFERKEY ET AL. amikacin and vancomycin for 33 isolates from bacteraemia and severe infections. MCs of methicillin were determined for 75 strains from many sources (wounds, burns, bacteraemia, serious infections and samples of air). The MCs of methicillin were 22 mg/l for all strains. Bacteriophage typing of GMRSA isolates The initial GMRSA isolate was sensitive to a wide range, including Group and Group 11 phages; it gave a type profile of 52/52A/6/42E/47/54/75/77/84/81/94 at 1 RTD. No S. aureus with this broad range was detected before or since. Subsequent isolates were of various phage types. Most isolates in 1976 and during the first 6 months of 1977 were type 77/84 or 77/+. Such isolates were sometimes found throughout the remainder of the study. GMRSA isolates of phage type 85 appeared in January 1977 and were the predominant type isolated during the second 6 months of 1977 and most of 1978. Strains that were untypable with the routine set of phages and some that gave inhibition reactions with a small number of group 11 phages at 1 RTD also emerged at the beginning of 1977. The latter isolates usually exhibited confluent lysis with the experimental phage 9 at 1 RTD. Most GMRSA isolates in 1979 and 198 gave this pattern. From 63% of isolates tested at various times (1978-198) were completely untypable. The phage types isolated from blood cultures and from other severe infection are shown in fig. 4. The different types found paralleled the overall phage typing pattern. No type was associated with bacteraemia more often than with severe infection without bacteraemia. The typability of all strains of S. aureus isolated in one hospital (MMH) was assessed over a 6-year period (1975-198). The results are shown in table V. Before 12-1 - - 6- BACTEAAEMA a r 1978 977 c9 1978 1979 18 6-8- 1-12 - u OTHER SEVERE NFECTON FG. 4. Phage types of GMRSA associated with bacteraemia and severe infections without bacteraemia. eo
RESSTANT S. AUREUS N DUBLN 125 TABLE V Typability of S. aureus isolates (M.M.H.), 1975-1982 Typing phages ~~~ Standard phage set Standard phages at RTD Standard phages at RTD x 1 Experimental phages Phage 9 Phage 88 Phage 89 Percentage of strains typable* with the given phages in,,. 1975 1976 1977 1978 1979 198 84 78 74.5 63 66 4.5 6 47 5 43 52 25 24 31 24.5 2 14 15-5.5 5 1 27.5 1 1 1 * Average number of strains typed per year = 3. Strains that were untypable with the standard phage set were tested with the experimental phages. the appearance of GMRSA in the hospital in 1978, ~74% of all S. aureus strains isolated were typed by the routine set of phages at RTD or RTD x 1. Subsequently the percentage typability decreased and, in 198,59-5% were untypable. Some of the untypable strains were lysed by an experimental phage (9) at RTD x 1 which was introduced into routine typing at RTD x 1 only in 1979. No antibiotic sensitive strains have been found that were lysed by phage 9; most were typable with the routine set of phages at RTD. DSCUSSON This survey shows that gentamicin and methicillin resistant S. aureus (GMRSA) have emerged as a serious problem both in the number and severity of infections in our hospitals. Most previously reported outbreaks of infection with gentamicin resistant S. aureus have been small and invasive infections rare. (Shanson et al., 1976; Speller et al., 1976; Public Health Laboratory Service, 1977; Wyatt et al., 1977; Faden et al., 1979; Price et al., 198). However, in a large outbreak of infection with strains of S. aureus resistant to methicillin and aminoglycosides, invasive infection was more frequently seen (Crossley et al., 1979). n contrast to the findings of Crossley et al. (1979), the multiply antibiotic resistant strains of S. aureus encountered in Dublin hospitals, in this study, were most frequently isolated from superficial sites. This experience was similar to that of other authors (Shanson et al., 1976; Speller et al., 1976; Warren and Roberts, 1976; Bint et al., 1977; Public Health Laboratory Service, 1977; Wyatt et al., 1977; Vogel et al., 1978; Price et al., 198). Bacteraemia and other severe infections were uncommon in the first 2 years of this study but were more frequently seen subsequently. This increase occurred as the organisms became more widespread in the hospitals. nfection or colonisation with GMRSA was not confined to these hospitals, but was an emerging problem in other Dublin hospitals throughout the study period (Hone et al., 198 1). Additionally, patients sometimes transferred from hospitals in different parts of the country were already colonised or infected with GMRSA on admission. Within the Dublin area, some services such as cardiac surgery, plastic surgery and neurosurgery, are located only in certain hospitals so
126 MARY T. CAFFERKEY ET AL. transfer of patients between hospitals is inevitable. Thus GMRSA may have been introduced or reintroduced into a hospital from any of these sources. Because of the wide dissemination of GMRSA throughout the rish Republic, it was interesting that, with the exception of one outbreak of gentamicin resistant infection (Wyatt et al., 1977), strains of S. aureus resistant to gentamicin were rarely isolated in Belfast hospitals throughout the study period (Wyatt, personal communication). t is not known whether these organisms are commonly found in other hospitals in Northern reland. n this study, as in the study of Greenhood et al. (1979), the predominant phage type changed with the progression of the outbreak. GMRSA isolates of different phage types were sometimes found in a single lesion, or in different sites in a patient, as had also been reported previously (Wyatt et al., 1977). Phage typing became less helpful as an epidemiological tool as the outbreak progressed. By the second half of 1978, most isolates were untypable with the routine set of phages, some of which typed only with the experimental phage 9 at 1 RTD. Such isolates were essentially untypable. n three other outbreaks, most strains were lysed by the experimental phage 9 in addition to other standard phages in groups and 111 (Speller et al., 1976; Bint et al., 1977; Wyatt et al., 1977). New phages or different typing methods are now needed to detect differences between strains in these hospitals. t is intended to investigate the possibility that serotyping and biotyping may detect differences between strains. We wish to thank Mrs Mary Foody for valuable assistance with preparation of the manuscript. REFERENCES ANNEAR, D.. 1968. The effect of temperature on resistance of Staphylococcus aureus to methicillin and some other antibiotics. Medical Journal of Australia, 1, 444-446. BNT, A. J., GEORGE, R. H., HEALNG, D. E., WSE, R. AND DAVES, M. 1977. An outbreak of infection caused by a gentamicin-resistant Staphylococcus aureus. Journal of Clinical Pathology, 3, 165-167. BLAR, J. E. AND WLLAMS, R. E.. 1961. Phage typing of staphylococci. Bulletin of the World Health Organization, 24, 771-784. CROSSLEY, K., LOESCH, D., LANDESMAN, B., MEAD, K., CHERN, M. AND STRATE, R. 1979. An outbreak of infections caused by strains of Staphylococcus aureus resistant to methicillin and aminoglycosides.. Clinical Studies. Journal of nfectious Diseases, 139, 273-279. FADEN, H., NETER, E., MCLAUGHLN, S. AND GACOA, G. 1979. Gentamicin-resistant Staphylococcus aureus: Emergence in an ntensive Care Nursery. Journal of the American Medical Association, 241, 143-145. GREENHOOD, G. P., HLL, D. L., DXON, R. E., CARTER, M. J. AND KANTO, W. P. 1979. Changing phage typing patterns of epidemic gentamicin-resistant Staphylococcus aureus. Evidence for transmission of gentamicin resistance. Lancet, 1, 289-29 1. HONE, R. AND KEANE, C. T. 1974. Characteristics of methicillin resistant Staphylococcus aureus. rish Journal of Medical Sciences, 143, 145-154. HONE, R., CAFFERKEY, M. T., KEANE, C. T., HARE-BARRY, M., MOORHOUSE, E., CARROLL, R., MARTN, F. AND RUDDY, R. 1981. Bacteraemia in Dublin due to gentamkin-resistant Staphylococcus aureus. Journal of Hospital nfection, 2, 119-126. LACEY, R. W. 1975. Antibiotic resistance plasmids of Staphylococcus aureus and their clinical importance. Bacteriological Reviews, 39, 1. LACEY, R. W. AND MTCHELL, A. A. B. 1969. Gentamicin-resistant Staphylococcus aureus. Lancet, 2, 1425-1426.
RESSTANT S. AUREUS N DUBLN 127 PRCE, E. H., BRAN, A. AND DCKSON, J. A. 198. An outbreak of infection with a gentamicin and methicillin-resistant Staphylococcus aureus in a neonatal unit. Journal of Hospital nfection, 1, 221-228. PUBLC HEALTH LABORATORY SERVCE. 1977. Unusual antibiotic-resistant staphylococcus in a hospital. British Medical Journal, 2, 325. SHANSON, D. C., KENST, J. G. AND DUKE, R. 1976. Outbreak of hospital infection with a strain of Staphylococcus aureus resistant to gentamicin and methicillin. Lancet, 2, 1347-1 348. SOUSSY, C. J., BOUANCHAUD, D. H., FOUACE, J., DUBLANCHET, A. AND DUVAL, J. 1975. A gentamicin resistance plasmid in Staphylococcus aureus. Annals of Microbiology (Paris), 126b, 91-94. SOUSSY, C. J., DUBLANCHET, A., CORMER, M., BSMUTH, R., MZON, F., CHARDON, H., DUVAL, J. AND FABAN, G. 1976. Nouvelle resistances plasmidiques de Staphylococcus aureus aux aminosides, (Gentamicine, tobramycin, amikacine). La Nouvelle Presse Me'dicale, 5, 2599-262. SPELLER, D. C. E. and 7 others. 1976. Epidemic infection by a gentamicin resistant Staphylococcus aureus in three hospitals. Lancet, 1,464-466. STOKES, E. J. AND WATERWORTH, P. M. 1972. Antibiotic sensitivity tests by diffusion methods. Association of Clinical Pathologists, Broadsheet No. 55. VOGEL, L., NATHAN, C., SWEENEY, H.M., KABNS, S. A. AND COHEN, S. 1978. nfections due to gentamicin-resistant Staphylococcus aureus strain in a nursery for neonatal infants. Antimicrobial Agents and Chemotherapy, 13,466-472. WARREN, R. E. AND ROBERTS, S.. B. 1976. Gentamicin-resistant staphylococci. Lancet, 2, 543-544. WYATT, T. D., FERGUSON, W. P., WLSON, T. S. AND MCCORMCK, E. 1977. Gentamicin-resistant Staphylococcus aureus associated with the use of topical gentamicin. Journal of Antimicrobial Chemotherapy, 3, 213-2 17.